Bert Vaux | University of Cambridge (original) (raw)

Uploads

Books by Bert Vaux

Papers by Bert Vaux

Journal of the International Phonetic Association, Mar 31, 2021

In this Illustration we describe the Cwyzhy (also Tswydzhy) dialect of Abkhaz, the native languag... more In this Illustration we describe the Cwyzhy (also Tswydzhy) dialect of Abkhaz, the native language of the third author. In Cwyzhy, the language Abkhaz is called /apHsaSWa/ [»apHsQSjQ] аҧсашәа. 1 Abkhaz (ISO-639-3 abk) belongs to the Northwest Caucasian family of languages, and the Abkhaz dialects are related as shown in (1) (adapted from Chirikba 2012: 36): (1) Relationships between Abkhaz varieties Cwyzhy is a sub-dialect under the Sadz node in (1). All of the language varieties dominated by the 'Proto-Abkhaz' node are mutually intelligible (Chirikba 1996b: ii). Though T'ap'anta and Ashkharywa are spoken by a politically distinct Abaza people who live in the North Caucasian republic of Karachaj-Cherkessia (ibid.: ii), linguists who work on Abkhaz (e.g.

EPL (Europhysics Letters), 2020

We introduce a stochastic model of language change in a population of speakers who are divided in... more We introduce a stochastic model of language change in a population of speakers who are divided into social or geographical groups. We assume that sequences of language changes are driven by the inference of grammatical rules from memorised linguistic patterns. These paths of inference are controlled by an inferability matrix which can be structured to model a wide range of linguistic change processes. The extent to which speakers are able to determine the dominant linguistic patterns in their speech community is captured by a temperature-like parameter. This can induce symmetry breaking phase transitions, where communities select one of two or more possible branches in the evolutionary tree of language. We use the model to investigate a grammatical change (the rise of the phrasal possessive) which took place in English and Continental North Germanic languages during the Middle Ages. Competing hypotheses regarding the sequences of precursor changes which allowed this to occur each generate a different structure of inference matrix. We show that the inference matrix of a "Norway Hypothesis" is consistent with Norwegian historical data, and because of the close relationships between these languages, we suggest that this hypothesis might explain similar changes in all of them.

Proceedings of NELS, 2001

Uyghur is generally believed to possess a vowel harmony system very similar to the one found in i... more Uyghur is generally believed to possess a vowel harmony system very similar to the one found in its relative Turkish, save for the fact that in Uyghur i is neutral and transparent (Lindblad 1990, Hahn 1991, Alling 1999). In this paper I argue on the basis of the phonological behavior of disharmonic vowels that Uyghur vowel harmony is actually quite different from the Turkish system in that harmony propagates only [-back] and harmony applies both cyclically and postcyclically. I demonstrate furthermore that the Uyghur facts ...

The Linguistic Review, 2015

We argue that the cross-linguistic distribution of vowel systems is best accounted for by grammar... more We argue that the cross-linguistic distribution of vowel systems is best accounted for by grammar-external forces of learnability operating in tandem with cognitive constraints on phonological computation, as argued for other phonological phenomena by

Revue des Études Arméniennes, 1996

Classical Armenian displays a curious type of subordinate clause formation characterized by the a... more Classical Armenian displays a curious type of subordinate clause formation characterized by the addition of a cliticized anaphoric pronoun to the end of the first constituent within the subordinate clause. Meillet (1897-1898) first proposed that this process was an instance of Wackernagel's Law, which encompasses a range of syntactic secondposition phenomena in the world's languages. In this paper I survey the distribution of Wackernagel clauses in Armenian, and provide an account for their behavior within the framework of current syntactic theory. More specifically, I show that the constraints on what elements may serve as host for the clitic and where the finite verb and adverbs in the subordinate clause may surface follow naturally from independently motivated principles of recent generative syntactic theory.

Proceedings from the Annual Meeting of the Chicago Linguistic Society, 2003

This paper explores English shm-reduplication and aims to answer questions that have previously b... more This paper explores English shm-reduplication and aims to answer questions that have previously been left open on the subject. Results from an online survey show emerging patterns that have not been adequately addressed in the literature. It will be shown that shm-reduplication targets prosodic landmarks, syllabic landmarks and the phrasal site of the target of reduplication. Instances of avoidance phenomena will also be discussed and analyzed. The results suggest that shm-reduplication is computed by a grammar that is ...

Linguistic Inquiry, 2000

Since Clements (1985) introduced feature geometry, four major innovations have been proposed: Uni... more Since Clements (1985) introduced feature geometry, four major innovations have been proposed: Unified Feature Theory, Vowel-Place Theory, Strict Locality, and Partial Spreading. We set out the problems that each innovation encounters and propose a new model of feature geometry and feature spreading that is not subject to these problems. Of the four innovations, the new model-Revised Articulator Theory (RAT)-keeps Partial Spreading, but rejects the rest. RAT also introduces a new type of unary feature-one for each articulator-to indicate that the articulator is the designated articulator of the segment.

Abstract: Patterns of plural selection in Armenian suggest that lexical representations of morphe... more Abstract: Patterns of plural selection in Armenian suggest that lexical representations of morphemes must include predictable syllabic structure, contrary to most theories of phonology, and that some phonological rules such as syllabification may precede morphological rules, contrary to the theory of distributed morphology. Furthermore, certain segments at the edges of morphological domains are not syllabified in lexical representations, and are syllabified at a later stage in the derivation. The findings are ...

Abstract The Western Armenian possessive plural data originally reported in Vaux (1998, 2003) hav... more Abstract The Western Armenian possessive plural data originally reported in Vaux (1998, 2003) have been asserted by Wolf 2011 to involve outwardly-sensitive phonologically conditioned allomorphy, a phenomenon widely argued to be unattested (Carstairs-McCarthy 1987; Paster 2006) and predicted to be impossible by the tenets of Distributed Morphology (Halle and Marantz 1993; Bobaljik 2000). We show that the full complexity of the Western Armenian system is better captured in an account that makes no reference to outwardly- ...

The Handbook of English Linguistics

IntroductionBenveniste 1952:" transitive perfect" nora e: gorUeal3Sg. GEN be. 3sg do-pp... more IntroductionBenveniste 1952:" transitive perfect" nora e: gorUeal3Sg. GEN be. 3sg do-ppl.'(s) he has done/accomplished'possessionnora e: handerdz3Sg. GEN be. 3sg garment'(s) he has a garment'TODAY'S POINTS: 1. The possession analysis of the transitive perfect was already known by the Armenianmonks in Vienna by 1866.2. More concrete evidence for the possession analysis occurs inKoriwn, Middle Armenian, and at least two modern dialects, which use lexical'have'to form the perfect with" transitive" verbs. 3. These ...

Journal of Indo-European Studies, 1992

Journal of the International Phonetic Association, Mar 31, 2021

EPL (Europhysics Letters), 2020

Proceedings of NELS, 2001

The Linguistic Review, 2015

Revue des Études Arméniennes, 1996

Proceedings from the Annual Meeting of the Chicago Linguistic Society, 2003

Linguistic Inquiry, 2000

The Handbook of English Linguistics

Journal of Indo-European Studies, 1992

Linguistic Society of America Annual Meeting, 2003

Phonologists generally assume that plain voiceless consonants are less marked than voiceless aspi... more Phonologists generally assume that plain voiceless consonants are less marked than voiceless aspirates, and that the unmarked two-way stop system contrasts unaspirated voiced and voiceless, as in Spanish; systems containing aspirates, as in English, are marked in comparison. I suggest that this picture is not entirely accurate. Building on a proposal by Steriade 1997, the maximally unmarked single-series stop is unspecified for laryngeal features, which is not the same as a voiceless unaspirated stop; the former may show wide significant variation, whereas the latter is specified for particular laryngeal gestures. I argue moreover, based on evidence from acquisition, articulation, perception, and L1 and L2 phonology, that the unmarked two-way stop system opposes aspirated and unaspirated stops, and that the aspirates may be the unmarked member of this set.

From Sounds to Structures

This paper is a contribution to a debate which has arisen concerning the necessity of readjustmen... more This paper is a contribution to a debate which has arisen concerning the necessity of readjustment, i.e. phonological processes sensitive to morphological information, as a device for implementing allomorphic alternations, the question being whether readjustment is necessary as such, or whether these alternations can be accounted for by a combination of listed allomorphs and regular phonology. We show that, given a piece-based morphological framework such as Distributed Morphology, an accurate description of the phonology of accent in Vedic Sanskrit requires readjustment or a close analogue; listed allomorphs do not suffice to capture the observed facts. We also discuss ways in which these readjustment rules may be constrained, observing that they require a degree of derivational articulation, and that they may not be readily segregated from those phonological processes lacking morphological conditioning.

Goldsmith/The Handbook of Phonological Theory, 2011

The Harvard Dialect Survey of 2002-3 represented the first linguistic foray into large-scale crow... more The Harvard Dialect Survey of 2002-3 represented the first linguistic foray into large-scale crowdsourcing (60K respondents) incentivized by dynamic geospatial imaging. Working in tandem with statistics graduate student Josh Katz of North Carolina State University I expanded this in 2013 to make the New York Times dialect quiz, which deployed Josh's brilliant tweaks of existing clustering, visualization, and prediction algorithms to attract responses to my survey questions from more than 21 million humans. Since that time I have been collaborating with forensic linguist Jack Grieve of Aston University to extract linguistically-significant patterns and trends from our megacorpus. In this talk I report on the development of the New York Times quiz and some of the leading discoveries that have emerged from it, including isogloss conspiracies and stability, the role of political and commuting zones, and multivariate non-local cultural regions.

An overview of work on age&language in connection with Sankoff and Labov's talks, together with a... more An overview of work on age&language in connection with Sankoff and Labov's talks, together with a bit of Tok Pisin.

This paper has two related goals, one armenological and one theoretical. My theoretical goal is t... more This paper has two related goals, one armenological and one theoretical. My theoretical goal is to provide a more accurate formal characterization of the syntactic, morphological, and semantic means employed to express possession and existence in the range of historical and regional varieties of Armenian. By showing that this can be done in a coherent manner that sheds light on the structure of Armenian in particular and human linguistic cognition in general, I hope as well to achieve my armenological goal of refuting Meillet’s claim that one cannot fruitfully study the syntax of Classical Armenian because it is calqued on Greek and/or Syriac.
My starting point is Benveniste’s well-known 1952 observation that one can make sense of the seemingly unusual Classical “transitive perfect” (e.g. nora ē gortseal ‘(s)he/it has done’, where the subject of (supposedly) transitive verbs is placed in the genitive case, whereas the subject of intransitives is nominative) by relating it to the Classical possessive construction, which also places the possessor in the genitive (e.g. nora ē handerdz ‘(s)he/it has a garment’). In Benveniste’s analysis, the transitive perfect is a possessive construction, parallel to the have-perfect of European languages such as French, Italian, German, Dutch, and (older stages of) English. I show that Benveniste’s generalization, which remains influential in the theoretical syntactic literature up to the present day (cf. Kayne 1993, Thibault and Sankoff 1997), requires two important modifications, one historical and one syntactic: (i) the possession analysis of the transitive perfect was already known by the Armenian monks in Vienna by the mid-nineteenth century (cf. Ajtənean 1866.2:96-97); (ii) these perfect constructions actually distinguish unaccusatives and passives (which select ‘be’) from transitives and unergatives (which select ‘have’). I then show that more concrete evidence for the possession analysis of the perfect occurs in Koriwn, Middle Armenian, and at least two modern dialects (Hamshen and Xotrĵur), which use lexical ‘have’ to form the perfect with “transitive” verbs.
Finally, I demonstrate that the distinction between two classes of intransitive verbs—unaccusative and unergative—shows up in several other situations in Armenian, including:

1. compound verb formation in many modern dialects, e.g. thəmaχ aniel ‘be greedy’ vs. mayil elniel ‘wonder, be amazed’ (Van dialect, Ačařyan 1952:189)
2. progressive aspect formation with guni in Hamshen (Ačařyan 1947:140-1)
3. formation of perfects of witnessed vs. unwitnessed events in some dialects (Grigoryan 1957:171-2)
4. formation of subject vs. object participles in nominal relative clauses in Standard Western Armenian (Sigler 1997)
5. selection of active vs. passive meaning with the -man participial suffix in the New Julfa dialect (Ačařyan 1940)
6. analytic vs. synthetic causative formation in Modern Armenian (Comrie 1981:182)

whistlestop tour of Armenia, Armenians, and the development of the Armenian language from Proto-I... more whistlestop tour of Armenia, Armenians, and the development of the Armenian language from Proto-Indo-European

Are the phonological generalisations formed by language learners upon exposure to underdetermined... more Are the phonological generalisations formed by language learners upon exposure to underdetermined data sets (i) maximally broad (dictated by maximal representational efficiency, as in rule-based phonology of the SPE tradition), (ii) maximally specific (dictated by the Subset Principle, as in Hale and Reiss's 2008 "The Phonological Enterprise", or (iii) variable in scope, depending on (a) minimisation of uncertainty about future events (Gallistel) or (b) Bayesian comparison of competing hypotheses (Tenenbaum)? In this talk I present evidence that a hybrid of (iiia-b) best accounts for what we know about phonological acquisition (and in fact animal learning as a whole), and circumvents problems encountered by theory (i) with scope variation and theory (ii) with trajectories of diachronic change and flaws in applying Gold's subset reasoning to phonology.

Optimologists generally consider Optimality Theory (OT) to be more constrained or at worst extens... more Optimologists generally consider Optimality Theory (OT) to be more constrained or at worst extensionally equivalent to Rule-Based Phonology (RBP) (q.v. McCarthy 1998, Mohanan 2000:145). In reality, OT conceals a number of profound pathological differences that have begun to be noticed in the literature, such as counterfeeding from the past (Wilson 2008), cascading credit problems (Pater 2005), and non-locality in Local Conjunction (Pater 2006). My ongoing Pandora's Toolbox project surveys a wide range of such cases where OT creates
unforeseen/unwanted consequences by combining all phonological and morphological processes into a single mega-operation. In the present talk I focus on the pathological consequences of separating structural configurations from repairs within a parallel processing framework. Contrary to optimologists' pou sto, the idea that OT but not RBP provides a unified and insightful account for phonological conspiracies, I show that:
(i) Single constraints are not in fact able to carry out the functions of rule conspiracies, due to a variant of the Underdetermination Problem. In fact, separating separating structural configurations from repairs entails that an infinite (or at least annoyingly large) constraint set is required to cover the territory of a single rule (pace McCarthy 1998 and Mohanan 2000).
(ii) Even if a single constraint could replicate a rule conspiracy, this would problematically (and in all likelihood wrongly) predict that tying that constraint to another relevant constraint in the hierarchy should produce an optional conspiracy.
(iii) OT also produces optionality cascades: tying two F constraints indexed to a set of M constraints causes all of these M constraints to change their repair simultaneously according to the ranking chosen for the tied F pair. (The same phenomenon would require several chained optional rules in RBP, a highly complex operation that appears to be unattested cross-linguistically.)
(iv) In addition to failing to account for the conspiracies it
identifies, OT creates a new host of conspiracies wherein a unified process such as English nasal place assimilation must for theory-internal reasons be attributed to several similar constraints (Idsardi 1997, Mohanan 2000).

Greeks and Armenians have been in continuous and extensive contact from Indo-European times to th... more Greeks and Armenians have been in continuous and extensive contact from Indo-European times to the present day. Until the fifth century, for instance, Christian worship in Armenia was conducted in Greek or Syriac. Greek influence on the Armenian language is well-documented in the biblical translations of the 5th century and the works of the philhellenic school of Armenian philosophers and translators (6th-8th centuries, though it is disputed who exactly belongs to this category). Greco-Armenian genetic resemblances predating the appearance of Armenian texts in the 5th century have been extensively investigated as well, notably in Clackson’s The Linguistic Relationship between Armenian and Greek (1994). Linguistic cross-fertilisation has continued into the modern era, as seen for example in the Greek dialects of Cappadocia and Pharasa in Turkey (Dawkins 1916) and Lori, Hankavan, and Yaghdan in Armenia (Hodgson 2008)
But what do we know about linguistic interactions between the Armenian and Greek cultures outside of this timeframe and beyond the pale of high literary culture? This talk provides a mise au point of the scattered available scholarship on the diversity of Greco-Armenian language contact in Late Antiquity and Byzantium and traces its early Christian antecedents and post-Byzantine consequences in the Balkan Sprachbund. Materials to be exploited include early Christian loanwords, Greek inscriptions in Armenia, Armenian inscriptions and manuscripts in Greek script, bilingual Byzantine seals and inscriptions, and Byzantine renditions of Armenian personal and place names.

Popular opinion maintains that World Englishes have been undergoing radical restructuring over th... more Popular opinion maintains that World Englishes have been undergoing radical restructuring over the past two decades or so due to drastic changes in the dissemination of American language and culture via films, television, and the internet. The same internet also happens to provide us with new tools. In this talk I survey what these new tools reveal about the present state of World Englishes, focusing on data collected from a series of large-scale online surveys I have run over the past seven years.

How can we tell how a language was pronounced long before we have live speakers, recordings, or e... more How can we tell how a language was pronounced long before we have live speakers, recordings, or even linguistically sophisticated descriptions? I address this question using the example of Armenian, an Indo-European language originally spoken in eastern Anatolia and the Caucasus and now spread in small numbers over most countries of the world. We will consider renditions of personal and place names in contemporary Byzantine documents, errors in manuscripts and inscriptions, phonological patterns in native and borrowed words, triangulation from modern dialects, and rendition of Armenian in non-Armenian scripts (Greek, Roman, Cyrillic, Arabic) and vice versa.

Hračhea Ačařean states in his 1913 Hayerēn Gawařakan Bařaran (1913:39) that Artvin "remains an en... more Hračhea Ačařean states in his 1913 Hayerēn Gawařakan Bařaran (1913:39) that Artvin "remains an entirely unknown dialect", and little has happened to change that picture in the subsequent century despite its putative membership in the mysterious lis dialect subgroup (e.g. ērthlis ēs 'you go'). The dialect is not mentioned in most surveys of Armenian dialects, and the lone thesis on the dialect (Alaverdyan 1969) is lost, according to directors of the dialect institute (personal communication, 2001). The most extensive treatment of the
dialect currently available is Gharibyan 1953:329-342, on which
Grigoryan 1957:435-452 and Alaverdyan (1968a, b) appear to be based. The original but limited presentations by Ačařean and Gharibyan are based on the speech of individuals who grew up in Artvin and Batumi before the Genocide, leaving us unsure of what has become of the dialect in subsequent generations. The primary purpose of the present talk is to address this lacuna based on field work I have conducted
with two middle-aged speakers of the dialect currently residing in the Los Angeles area. I focus on the following questions:
• Where is the dialect currently preserved, and by how many speakers?
• How has the dialect changed from the form reported by Ačařean, Gharibyan, and Alaverdyan?
• How does the dialect related to the neighboring Armenian dialect groups (Hamshen, Trabzon, Xodrĵur, Erzerum/Axaltsxa/Axalkalak, Tiflis) and the other dialects of the -lis branch?

If you have an iphone and speak american english (or enjoy pretending to), try out my new US Dial... more If you have an iphone and speak american english (or enjoy pretending to), try out my new US Dialect app! The idea for this app occurred to me about 4 years ago when visiting my old student Jim Houghton in London, and he showed me a 20 Questions app on his phone. I immediately thought that the same sort of thing could/should be done for English dialects using the data from my old Harvard Dialect Survey with Scott Golder . Sadly I couldn't find anyone with programming skills to go in on it with me until Adrian Leemann , Daniel Wanitsch , and the rest of the team in Zurich fortuitously showed up 1.5 years ago, and now our scheme has finally come to fruition.
The more respondents we get the more accurate the prediction will become, so you'll be greatly helping me out if you get all of your friends and family to try it!

Investigates the claim that epenthesis cannot counterbleed voicing assimilation.